I was speaking to Marc Doty earlier about it. He said that it’s not actually bi timbral. It’s a weird configuration where the parameters of the synth actually change between the patches rather than having two separate engines. He said he never noticed any issues with polyphony. He said with the morphing and the effects you nevertheless notice any note stealing.
To be honest I think 12 to 16 voices on a synth like this would sound awful given the characteristics of the Brute style sound. It would just sound like distorted mud. I think why Sequential synths work well as poly synths is because the oscillators themselves have a bit of a thin sounding character to them but when played against each other it has a pleasing sound....similar to Orchestral brass or strings. The problem with having oscillators that sound massive with a single note is it starts to become too cacophonous. Like a bottleneck effect. I’d you think of brass sections...if you have a total of six French horns it produces a powerful sound but it’s still distinguishable...you can hear each note...now say you have 100 French horns or 200..the natural phasing starts to become overpowering and it starts to become indistinguishable. Too much of a good thing.
Interesting... thanks for the tip on bitimbral - just looked this up in the manual. There is definitely a Layer/Stack mode, but it is unique in its implementation:
Pg 38:
Layer: the keyboard has one zone, with two morphing voices per note. If the Morph control is at zero, sound A is doubled. When it is at maximum, sounds A + B are doubled.
So, in one state, sounds like it acts as a sort of Poly-2 Unison mode, in the other state its a poor man's bi-timbral stack architecture.
The limitation would be that each of the two "layers" in the stack must share the same mod matrix routing arrangement. I suppose on the A side you could have a bunch of the mod routings just set to 0 value, and same on the B side with non used modulations at 0, which would essentially allow different effective mod routings per layer, but definitely not as straightforward as just designing two different timbres from scratch... I would expect from a sound design standpoint it will still be possible to pull off a great majority of bi-timbral stack type of sounds though. (But with the very low, 3-Voice Stack limitation)
Regarding more voices: I don't really think Sequential oscillators sound thin... Also, as an owner of MatrixBrute I can say that its quite capable of pulling off extremely subtle, beautiful sounds - soft leads and paraphonic pad type of sounds (same is true of the Pro 3)... and the PolyBrute, with the same sort of architecture should be capable of creating a versatile range of sounds from aggressive / brutish sounds to subtle and smooth tones. It may have "Brute" in its name, but the filters and overall sound is extremely versatile.
I would imagine there would only be bonus to having 12 voices instead of six available... and it would just open up the capability for more cinematic type of sounds and other long release patches. Will have to wait until November to tell for sure